Amir 

import pandas as pd

import numpy as np

import matplotlib.pyplot as plt

from scipy.stats import gaussian_kde

from sklearn.cross_validation import train_test_split

import pandas_highcharts

khodro = pd.read_csv('khodro.csv')

khodro.columns = ['name', 'open', 'high', 'low', 'close', 'Today Volume', 'Yesterday Volume', '2 Days ago Vol',

         'Individuals Ratio', 'Individuals Sell Ratio', 'Trade Volume', 'Month', 'Day of Month',

         'Day of Week',

         'Latent Variable']

khodro['pdif'] = khodro['close'] - khodro['close'].shift(1)

# print khodro.tail()

khodro['profit'] = khodro['pdif'].apply(lambda x: 0 if x < 0 else 1)

khodro['close'] = (khodro['close'] - khodro['close'].mean()) / khodro['close'].std()

khodro['Today Volume'] = (khodro['Today Volume'] - khodro['Today Volume'].mean()) / khodro['Today Volume'].std()

khodro['Yesterday Volume'] = (khodro['Yesterday Volume'] - khodro['Yesterday Volume'].mean()) / khodro[

  'Yesterday Volume'].std()

khodro['Individuals Ratio'] = (khodro['Individuals Ratio'] - khodro['Individuals Ratio'].mean()) / khodro[

  'Individuals Ratio'].std()

x = khodro[['close', 'Today Volume', 'Yesterday Volume', 'Individuals Ratio']]

y = khodro['profit']

x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)

x1 = x_train[y_train == 1]

y1 = y_train[y_train == 1]

x0 = x_train[y_train == 0]

y0 = y_train[y_train == 0]

xs = np.linspace(-2, 3, 200)

##Close-Plots

density1 = gaussian_kde(x1['close'].values)

density0 = gaussian_kde(x0['close'].values)

plt.plot(xs, density0(xs), c='red')

plt.plot(xs, density1(xs), c='green')

plt.hist(x0['close'], 52, normed=1, alpha=0.7, color='red', label='close-loss')

plt.hist(x1['close'], 52, normed=1, alpha=0.7, color='green', label='close-win')

plt.scatter(x0['close'], y0.values, c='red')

plt.scatter(x1['close'], y1.values, c='green')

plt.legend()

plt.show()